Chemical fingerprinting of unevaporated automotive gasoline samples

The comparison of two or more samples of liquid gasoline (petrol) to establish a common origin is a difficult problem in the forensic investigation of arsons and suspicious fires. A total of 35 randomly collected samples of unevaporated gasoline, covering three different grades (regular unleaded, pr...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Forensic science international 2003-06, Vol.134 (1), p.1-10
Hauptverfasser: Sandercock, P.M.L, Du Pasquier, E
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:The comparison of two or more samples of liquid gasoline (petrol) to establish a common origin is a difficult problem in the forensic investigation of arsons and suspicious fires. A total of 35 randomly collected samples of unevaporated gasoline, covering three different grades (regular unleaded, premium unleaded and lead replacement), were examined. The high-boiling fraction of the gasoline was targeted with a view to apply the techniques described herein to evaporated gasoline samples in the future. A novel micro solid phase extraction (SPE) technique using activated alumina was developed to isolate the polar compounds and the polycyclic aromatic hydrocarbons (PAHs) from a 200 μl sample of gasoline. Samples were analysed using full-scan gas chromatography–mass spectrometry (GC–MS) and potential target compounds identified. Samples were then re-analysed directly, without prior treatment, using GC–MS in selected ion monitoring (SIM) mode for target compounds that exhibited variation between gasoline samples. Principal component analysis (PCA) was applied to the chromatographic data. The first two principal components (PCs) accounted for 91.5% of the variation in the data. Linear discriminant analysis (LDA) performed on the PCA results showed that the 35 samples tested could be classified into 32 different groups.
ISSN:0379-0738
1872-6283
DOI:10.1016/S0379-0738(03)00081-1